Computer science researchers have come up with a method that achieves results comparable to expert-grade polysomnography using only a single-lead echocardiogram. This breakthrough not only simplifies a process that is usually expensive and impractical but also reveals a deeper link between the heart and the brain than previously recognized.
Quality sleep is crucial for survival, just like food and water. Despite spending around a third of our lives asleep, the nature of sleep remains largely a scientific enigma.
However, experts haven’t been idle in attempting to understand it.
Polysomnography, or sleep analysis, is employed to diagnose sleep disorders by collecting various types of information, including brain activity (from electroencephalograms or EEG) and heart function (via electrocardiograms or ECG). This often involves patients being connected to multiple sensors and wires while they sleep in a clinic, which doesn’t make for the most relaxing environment.
But what if the same test could be performed at home, with equal accuracy and in real time?
For the first time, researchers at the University of Southern California have developed a method that rivals expert-level polysomnography using just a single-lead echocardiogram. The open-source software enables anyone with basic coding skills to create an affordable, DIY sleep-monitoring device.
“For years, researchers have sought simpler and cheaper ways to monitor sleep without the cumbersome equipment,” said the lead author, Adam Jones, who recently earned his PhD from USC. “Historically, poor performance, even under ideal circumstances, suggested that it was impossible, and that monitoring brain activity was essential. Our study demonstrates that this assumption is outdated.”
The model, which accurately identifies sleep stages at the highest level, also outperformed other models that don’t use EEG, as well as commercial sleep-monitoring devices, noted the researchers. “We set out to create a system that addresses the current methods’ limitations and meets the demand for greater accessibility and affordability in sleep analysis,” explained Jones.
The study, published in June 2024 in the journal Computers in Biology and Medicine, was co-authored by Laurent Itti, a computer science professor and Jones’ advisor, along with Bhavin R. Sheth, a USC alumnus and electrical engineer at the University of Houston, who has collaborated with Jones for many years.
Is the heart leading the way?
Sleep is a significant indicator of cognitive decline, which tends to become shorter and more disrupted as we age—a trend confirmed by past studies and the researchers’ neural network. Interestingly, this decline can begin earlier than anticipated. Research published in Neurology revealed that individuals experiencing more disrupted sleep in their 30s and 40s are more than twice as likely to face memory issues a decade later.
Additionally, ongoing poor sleep may contribute to the buildup of beta-amyloid plaques, a key indicator of Alzheimer’s disease.
“It’s a bit alarming,” admitted Jones, who previously belonged to the “sleep when I’m dead” mindset before diving into sleep research as a hobby in 2010. “That’s why I’m eager for these interventions to be implemented quickly and made accessible to everyone. This software could help uncover what happens during our nightly sleep.”
The researchers trained their model using a broad and diverse dataset comprising 4,000 recordings from subjects aged between 5 and 90, relying solely on heart data and a deep-learning neural network. Through extensive experimentation, they determined that their automated ECG-only system could evaluate sleep just as effectively as the traditional “gold standard” polysomnography. This model successfully distinguished all five sleep stages, including rapid eye movement (REM) sleep—crucial for memory retention and emotional balance—and non-REM sleep, particularly deep sleep, which is vital for physical and psychological rejuvenation.
This innovation not only simplifies a typically costly and complex procedure but also suggests a closer connection between the heart and the brain than previously known. It emphasizes the role of the autonomic nervous system, which connects the two organs.
“The bond between the heart and brain is complex and not entirely understood, and this research seeks to bridge that knowledge gap,” stated Jones. “There’s substantial evidence in our findings that the heart might actually be playing a leading role.”
This work could also enhance sleep studies in remote communities, providing insights into the purpose and origins of sleep.
In a follow-up study currently being compiled, Jones plans to delve deeper into what the network reveals through the ECG data. “I believe there is a wealth of information within the heart that remains unexplored,” he remarked.
